Models/K-EXAONE

MDL

K-EXAONE

LG AI Research

Verified against Artificial Analysis · Jul 27, 2026

Capability

percentile index

Price /M

$0 / $0

in / out

Context

—

Avg score

59.5

10 benchmarks

AA Intelligence

22.1

index

Weights

Closed

Released

Dec 31, 2025

About

K-EXAONE is a proprietary model in the K family from LG AI Research. Benchmarked on 10 evals, averaging 59.5.

Benchmark scores · 10

avg 59.5 — higher is better

AIMEartificial-analysis · 2026-07

90.3

MMLU-Proartificial-analysis · 2026-07

83.8

GPQA Diamondartificial-analysis · 2026-07

78.3

LiveCodeBenchartificial-analysis · 2026-07

76.8

τ²-benchartificial-analysis · 2026-07

74.3

IFBenchartificial-analysis · 2026-07

64.7

AA Long-Context Reasoningartificial-analysis · 2026-07

55.7

SciCodeartificial-analysis · 2026-07

35.6

Terminal-Bench Hardartificial-analysis · 2026-07

22.7

Humanity's Last Examartificial-analysis · 2026-07

13.1

Details

PublisherLG AI Research

FamilyK

Parameters—

License—

WeightsClosed

ReleasedDec 31, 2025

Performance & price

Price in / out$0 / $0 per 1M

Coding index32.1

Math index90.3

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$0/M